Stueyb

Right, where to begin...

I have several dell nt 4 sp6 machines sitting on an internal network lan on the 10.0.x.x range.

Ther problem is that not all servers can see each other, but can ping each other no problem. For example ntserver4 (10.0.0.7) cannot talk to Chess2 (10.0.0.133) It can however ping it. Chess 2 can also talk to ntserver4. They are all members of the same domain also and there are no account sync problems on the network (that I know of)

Any ideas, as this is driving me up the wall.

Regards
Stuart

Jeff Wiltshire

It could be any number of things

Name Resolution (DNS or WINS)
Subnet mask
ARP

Are they all on the same switch ? Can you ping by name & IP addresses.....etc

Stueyb

I can indeed ping by address. The subnet mask for both machines is 255.255.0.0. Both machines are on the same switch.

Tony_W

Have you tried turning the hubs/switches/routers off and back on again?

I had a similar problem the other day, and this fixed it.

Tony

Stueyb

Ill give it a go and see what happens. Fingers crossed.

rich101

I agree with Jeff -

Host name resolution issue

Either assign one of the servers as the WINS or DNS server and set that in the TCP/IP settings of all of them ensuring the designated WINS or DNS server is started 5 minutes before the rest of them.

search on Google for "WINS nbtstat ping" should help with the finer detail.

rich

ChristianR

iTrader: (1)

what are the machines using as the default gateway?

Stueyb

The default gateway is 10.0.0.254. Looking the eventlog and it mentions Device\nwipxlnk is failing. Will lookup the nbstat pings and see whats what

workshy_fopp

Sound like wins is wrong.

Put in a hosts file as a temporary measure to prove if its name resolution or not.

dan4

If you have a wins issue you need an lmhosts file. try and ping by netbios name, if you can't add WINS service to one of your servers, and ammend network settings of all clients to use it, or add an lmhosts files (the sample file will help you construct it correctly). Good luck

David_Wallis

default gateway shouldnt matter as you are on the same subnet..

Check the bindings..

Wins shouldnt matter either.. they should talk without wins dns..

try doing start - run - \\10.0.0.1\c$ or whatever the ipadress is..


David

Foot_Tapper

Must admit, i think its a wins problem
hmmmmmmmmmmmmm
i would add an entry in the hosts file on chess2 for :
10.0.0.7 ntserver4
and see what happens.

David_Wallis

surely it wants to be in lmhosts.

if its a wins issue.. try this..

Get a hub build three pcs give them ip adresses 10.0.0.1 10.0.0.2 10.0.0.3 and a mask of anything you like within reason 255.255.255.0

allbeing well (microsoft ) you will be able to browset the machines in network neighbourhood, share files and connect to each other..

With no wins server.. hence not wins related more netbios related.

Are you filtering ports on your switches?? running layer 3 / 4 routing??

I'm 99% sure its nothing to do with a wins server or a dns server

what do you get as node type if you do ipconfig /all

is it a hybrid node?

David

rich101

>I can indeed ping by address. The subnet mask for both machines is 255.255.0.0. Both machines are on the same switch

So basic IP connectivity is there ruling out gateways, subnets etc

What protocols are currently enabled ? TCP check the bindings order and make sure TCP is at the top of the list.

If that is all OK then lmhost file are the only way forward, you can use a single file which lists each server and copy it to all as found here c:\$winnt$\system32\drivers\etc\lmhost use the existing example to add your servers

rich

RB5#295

That suggests IPX/SPX is the problem. Is Netware in the equation?

I agree it sounds like WINS at first view, but two machines on the same subnet and switch that can ping by IP can resolve by broadcast, so that shouldn't be it.

I'd try rebooting everything to see if it just goes away. If that fails I'd look up that device failure in Technet / Google (groups is very good if you haven't tried it before), assuming it has appeared for the first time around the time the problem appeared.

Fen (who can't believe he's on here rather than troubleshoot servers and now is troubleshooting servers )

RB5#295

Have you run some nbtstat tests?
nbtstat -r to see how each machine is resolving.
nbtstat -c to see if the name is cached to the wrong address (if at all)
nbtstat -R to flush the cache

You could also try ipconfig /flushdns to flush any DNS cache errors.

I had a cluster a couple of weeks ago where clients resolved to the fixed IP address of a node and when the resource failed over they still tried that address and got 'resource not found' or something.

Fen

Wurzel

iTrader: (1)

It may have absolutely nothing to do with it but why are you using 10.x.x.x? this is a reserved address range along the lines of the 127.0.0.1 address, I seem to remember having comms problems on some machines we set up using the 10.x.x.x ip address.

Jeff Wiltshire

10.x.x.x is part of the RFC 1918 private addresses and are reserved for 'Private' networks. 127.0.0.1 is reserved for loopback.

Not the same thing really.


Jeff

RB5#295

Not any more.

Cisco (IIRC - although I don't understand why it's up to them) have decreed that it is now a registered range. The NHS are already using it. Some sites also have it as private networks. The fun will begin soon when one of them wants to route to a new 10.x.x.x address across NHSNet...

Why shouldn't StueyB use it anyhow? I assumed it was his private scope behind NAT. That said I'd look at re-ip-ing pretty soon with 10.x.x.x now being registered...

Fen

ozzy

OK, I'll add in my 2p worth.

What is the error message you are receiving and is there any errors in the Event logs?

Try creating a HOSTS file with ALL the servers and their corresponding IP addresses listed. Then just copy that to ALL the servers \winnt\system32\drivers\etc.

Drop to a command prompt and try pinging each servers name i.e. ping nt4server OR ping chess2. That should work fine since you've already proved basic IP is working.

Next, try connecting to each servers hidden share (if you haven't removed them) i.e. \\nt4server\c$ or \\chess2\c$. That'll prove that name resolution is working AND authentication is working as it'll try to authenticate as the current logged-in user (I'd suggest the domain admin to rule out permissions issues).

If all the servers have TCP/IP install, are all in the same subnet and members of the same domain you don't need WINS, HOSTS files, LMHOSTS files, DNS or anything like that unless ....

... check you have Computer Browser, Workstation and NetBIOS Interface installed on each of the servers. If those aren't installed or starting you'll need higher-level protocols like WINS, DNS, etc..

Stefan

ChrisB

Correct.

I have two PCs plugged into a hud here. No WINS, only DNS is ISP provided and I can browse shares on the other Win98SE PC from my Windows 2000 PC.

I missed reading about 10.x.x.x becoming public - time to hit Google!

Jeff Wiltshire

RB5

Have you any documentation that says that RFC1918 is now superceeded ????

I'd be very surprised if 10.0.0.0/8 is no longer a Private Address range. Even if the NHS use it for their internal addresses doesn't mean it's status has changed.

IP Block record from IANA

OrgName: Internet Assigned Numbers Authority
OrgID: IANA
Address: 4676 Admiralty Way, Suite 330
City: Marina del Rey
StateProv: CA
PostalCode: 90292-6695
Country: US

NetRange: 10.0.0.0 - 10.255.255.255
CIDR: 10.0.0.0/8
NetName: RESERVED-10
NetHandle: NET-10-0-0-0-1
Parent:
NetType: IANA Special Use
NameServer: BLACKHOLE-1.IANA.ORG
NameServer: BLACKHOLE-2.IANA.ORG
Comment: This block is reserved for special purposes.
Comment: Please see RFC 1918 for additional information.
Comment:
RegDate:
Updated: 2002-09-12


[Edited by Jeff Wiltshire - 2/20/2003 10:00:12 PM]

ozzy

I can't see any of the 10.0.0.0, 172.16.0.0 or 192.168.0.0 addresses becoming public as billions of people must be using them for internal addresses.

I know Sky TV use these internally on all their systems, so just coz the NHS use them doesn't necessarily mean it's went public.

Stefan

dsmith

Indeed the NHSIA have recently decreed that new deployments within NHSNet are to use Network 10 - allocations of which will be controlled centrally. This is not to suggest its now public - precisley the opposite, its to be used *because* its private. The NHS - like most large companies/organisations - cannot just sufficicent RIPE allocations to satisfy all their internal needs - hence they use RFC1918 addresses. GPs have used RFC1918 172. addresses for the past few years. the NHSIA recommend 192.168 addressing on small networks wich hide behind NAT firewalls/routers before hitting NHSNet (some large GPs use 192.168 addresses locally then NAT to their small assigned range.

NHSNet itself is a private intranet and allows no direct access *from* the Internet. Both halves (provided by BT and C&W) provide a firewalled access *to* the Internet which are entirely NATted. For any application which needs to be accessed from both within NHSNet and the Internet there are stringent security assesments to ensure the NHSIAs security principles are not breached.

Its because the NHS Addresses (Both RIPE & RFC1918) are hidden from the Internet that the NHS would struggle to justify sufficient new RIPE allocations to meet their complete needs.

The use of Network 10 "publicly" within NHSNet *is* going to cause some of the larger NHS organisation which have alreeady deployed it internally some problems, however the NHSIA has set the strategy (and it the only sensible one to cope with the forecast growth) and they have plenty of time to cope with it.

see here for the NHSIA faq.

Deano

[Edited by dsmith - 2/21/2003 8:51:38 AM]

SiCotty

RB5#295,

I would very much like to know where you came across this information because as far as I am aware the 10.0.0.0 address range will never be routed across the public Internet.

The biggest waste of space is the 127.0.0.0 address range which can only be used for internal loopback addressing.

Si

RB5#295

Fair comment above; it could well just be within NHSNet, which makes more sense.

I was told this by someone else in the office and I have seen it in practice within NHSNet, but since that's technically just a very big Intranet...
The decision was attributed to Cisco, which I always found odd, but would clearly apply to the internet as a whole...

Anyhow apologies if I misled anyone - I have neither the time nor the inclination to check all this sort of stuff that I get passed on to me, particularly in this instance as I don't ever choose network scopes, just work with what's already in place.

Fen

dsmith

I could go on at some length about NHS addressing policy, particularly the allocation of RIPE addresses between the BT and C&W halves of NHSNet and the complete lack of summarisation that is possible. Which is what makes it impossible to ever seriously consider NHSnet being a stanadalone AS attached as part of the true Internet and why it will remain a private intranet behind NAT firewalls.

But I wont, as it keeps me in Scoobies

Deano

RB5#295

No offence, but I'm not that fussed anyway Deano, the sooner I'm not involved with NHSNet the happier I'll be.

Who do you work for then, the IA?

Fen

[Edited by RB5#295 - 2/21/2003 6:14:04 PM]

dsmith

Fen - no offence taken ???

I work for BT on BTHealthNet and spend my life trying get around the many and varied historical imperfections within it.

Anything in particular on NHSNet (or the HealthNet side anyway) causing you grief ?

Deano

[Edited by dsmith - 2/21/2003 7:48:01 PM]

RB5#295

Deano,

Nothing in particular giving me grief, no, I'm just completely fed up with my job - email servers in the NHS mainly, with a fair bit of AD thrown in (I work for BDS if that means anything to you).

We get a lot more hassle from the C&W sites than your's!

How does the plans N3 affect you guys?

Fen